NEWS

ces 1.1.0

Added the 2025 Canadian Election Study web survey, retrieved from Harvard Dataverse (doi:10.7910/DVN/2YAQ8D). Access it with get_ces("2025") or get_ces("2025", variant = "web"); the PDF technical report and codebook are available via download_pdf_codebook("2025").
Dataset coverage now spans 23 datasets from 1965 to 2025.

Restored access to the 2015 phone and 2015 combined (web/phone) surveys, which had stopped downloading after the original UBC host began returning an HTML bot-protection page to programmatic clients. The 2015 combined survey is now sourced from the Borealis repository, and the 2015 phone survey from a stable mirror, so both download reliably. Downloads also send a standard browser User-Agent (saved and restored on exit) for robustness against such host-side protections. Both datasets match the corresponding cesR package variants exactly (phone: 4,202 x 455, Latin1 encoding; combined: 11,614 x 665).

MAJOR: Fixed critical bug in create_codebook() function that caused "arguments imply differing number of rows" error for multiple datasets (2015 web, 2015 phone, 2021 web). The issue was caused by multi-value variable labels being incorrectly handled, leading to vector length mismatches during data.frame creation. Multi-value labels are now properly collapsed with " | " separator.
Added comprehensive tests for create_codebook() function to prevent regression, including specific tests for multi-value labels, missing labels, and different output formats.
Verified fix works across all CES datasets and maintains backward compatibility.

Fixed critical bug in 2015 combo dataset where file format was incorrectly specified as SPSS instead of Stata format, preventing get_ces("2015", variant = "combo") from working properly

Complete Survey Variant Support: Added comprehensive support for all CES survey variants (web, phone, combo, panel studies)
Enhanced Data Coverage: 22 CES datasets across 16 election years (1965-2021) with 100% codebook coverage
Improved User Experience: Intelligent defaults and informative messaging for variant selection
Robust Download System: Enhanced error handling, retry mechanisms, and ZIP file support
Comprehensive Testing: 65 passing tests with 0 failures

Initial release
Added functions for accessing Canadian Election Study datasets:
- get_ces() - Download and load CES data for a specific year
- list_ces_datasets() - List available CES datasets
- get_ces_subset() - Get a subset of variables from a CES dataset
- create_codebook() - Generate a comprehensive codebook for CES datasets
- export_codebook() - Export codebooks to CSV or Excel formats
- examine_metadata() - Analyze metadata across variables in a dataset
- download_pdf_codebook() - Download official PDF codebooks for any CES year
- download_ces_dataset() - Download a single CES dataset by year
- download_all_ces_datasets() - Download all or selected CES datasets at once
Enhanced metadata preservation:
- Metadata preservation is now enabled by default
- All variable labels and value labels are automatically maintained
- Complete preservation of original dataset attributes for better data documentation
Supported datasets from 1965 to 2021
Added vignette and examples
Included support for variable documentation and metadata exploration